One day after our morning meditation my friend Mark - an American living in Prague - said to me - what if we built a language app? I just replied “Let’s do it”. Little I knew this moment will change the course of my life for years to come.
Today, I’d like to present to you Latudio: https://www.latudio.com - a language acquisition app that helps you learn languages naturally through listening. Currently, we support French, Czech, Latin American Spanish, and Italian (all paired with English), with Brazilian Portuguese coming soon (available in preview at the moment).
The problem with most language apps is they focus heavily on reading and writing from day one. With our app we took a slightly different approach to acquiring a language, a “hack” in a way. All exercises are for listening, there is no text to read right from the start. But, if you’re ever lost and don’t understand what’s being said, you can pause the audio anytime to see the sentence. Any word in the sentence can be tapped on and you will see its contextual translation, which may be sometimes different to a dictionary translation of the word. And, each word tapped gets saved in your dictionary for later reviewing. In a dictionary, you can list all sentences containing the word in all various forms it can take - a great help if you want to explore words and get the feeling of the language.
The app contains several types of exercises:
Sentences - listen, reorder words, write the sentence, then record it.
Stories - listen to the story, then answer questions to be sure that you understood well. Also you can match sentences to its correct photo.
Conversations - scripted conversations where you take part as one speaker and try to record your voice to match the original recording.
Booster - like Sentences exercise, but you can select your own set of sentences.
You can also set goals and view statistics about your progress.
None of this would be possible without the great help from hundreds of our contributors [1], who translated, mapped and recorded content. All the content you find in the app was reviewed multiple times by several people and recordings are made by native speakers.
No story in the app was created by a LLM, all exercises are created and curated by us.
Privacy-wise, we collect as little information about our users as possible. For example, we didn’t include Facebook SDK even if it would be easier for us to implement Facebook login that way. Deleting your account really makes our app forget everything about you.
Some technical details - the app is written in React Native, API is written in Java (Javalin [2]). Our content development platform uses Apache Wicket [3]. BunnyCDN is used for content delivery. "Use boring technology", they say, so we did.
We rent a bare metal server from Hetzner and use Proxmox for VMs management. Our setup is fairly simple but flexible enough for us to grow if we need to. My approach always was, that if we’d need to figure out how to scale beyond VMs and multiple servers, we’d also have finances to hire a DevOps person for that. It will be a nice problem to have.
Latudio is completely bootstrapped, without any VC funding, allowing us to focus on our vision and users' needs.
Hope you’ll like the app. If you’d like to know more about our journey, I’ll be happy to answer your questions in the comments.
P.S.: we offer the app for free to nonprofits who take care of people that are at a disadvantage. We can also create custom content tailored to their specific needs - like training exercises for hospital workers or language exam prep for refugees. If you know someone who could benefit from this, we'll be happy to know. E-mail is in my profile.